One of the areas I focus on in my job is the social behaviors we enable and then capture in digital spaces, such as tagging, rating, commenting, sharing (email) etc. I am particularly interested when I see how other companies leverage similar capabilities in their products.

Adobe Photoshop Lightroom offers hierarchical classification (might have been in their Album product, I don’t remember), wherein the keyword Animals might contain Dog and Cat. Tagging an image with Dog or Cat implicitly adds Animals. However, if I had another keyword Family and add Dog as a child tag, Lightroom essentially sees two Dog tags and will only see the implicit tags if specified explicitly. If all I type in is Dog, it picks Dog > Animals and I do not benefit from the Dog > Family relationship.

From an end user point of view, this is when a flat list of tags is often more powerful, there are no semantics between tags upon entry, users just type away. This brings me to one of the elements I am curious about, the post processing of tags to understand their semantic relationships so that Dog can be both a part of Family and Animals without the user worrying about declaring either. The notion of implicit tagging is interesting, but it is not implicit if a user needs to take an additional action to experience the benefit.